California Democrat Ted Lieu is something unusual in Congress: He was a computer science major at Stanford, a course of study that led him logically to his interest in the issue of artificial intelligence, where he’s sponsored a number of bills and become one of the House’s leading experts.
A.I., as Lieu notes, has been around for a long time. But it’s generative A.I.—the kind that uses algorithms to create new content in video and other platforms—that Lieu says we need to be worried about. With the old A.I., he says, if you asked it what a cat was and showed it photos of a cat and a dog, it could pick out the cat. Today’s A.I.? “Now it generates 26 images of cats and writes you an essay about cats,” he said.
Certain potential concerns are obvious. “There are algorithms now that if you give like half an hour of a voice recording, it can mimic, basically, that person,” Lieu says. “Give it six hours’ worth, and it will sound exactly like them.” Hello there, audio or video of “Joe Biden” laughing about taking bribes from those gullible Ukrainians.
Lieu, who is also one of only two Taiwanese members of Congress (the other being Grace Meng), expressed confidence that people can be adequately forewarned to be on the lookout for such fakery. He also noted that there are some aspects of the new A.I. that we all can be excited about—it can help in diagnoses of diseases, for example. But making sure we get more of the good than the bad will require the federal government to step in, which is why he’s proposed (among other measures) that the government create a commission, with six members each appointed by the Senate and House and eight by the president, to issue a set of recommendations for what the government should do within two years. We just have to make it through the next election.